How to auto-tag images
Manually tagging hundreds of product photos is slow and inconsistent. Auto-tagging lets PixelFiddler’s AI examine each image, identify what it sees, and attach descriptive, searchable tags — so you can find any product photo in seconds instead of scrolling through folders.
This page explains how the AI works, what it detects, and how to get the most accurate results. If you are looking for guidance on organizing and managing your tags at scale, see the tag management workflow page instead.
What auto-tagging does
Section titled “What auto-tagging does”When you run auto-tagging on an image, the AI scans the photo and generates a set of keyword tags that describe its contents. These tags are saved directly on the image and become searchable across your entire library.
For example, a photo of a red leather handbag on a white background might receive the tags: bag, handbag, red, leather, white background, studio.
Once tagged, you can use the search bar to type any of those words and instantly find every matching image in your collection.
What the AI detects
Section titled “What the AI detects”The AI is trained to recognize the types of detail that matter most in e-commerce photography. It looks for five categories of information in every image.
Object types
Section titled “Object types”The AI identifies the main product in the photo and tags it with specific names. Examples include dress, bag, shoe, watch, sunglasses, necklace, chair, and lamp. It also detects secondary objects that appear alongside the main product.
Colors
Section titled “Colors”Dominant and accent colors are detected and tagged. You will see tags such as red, blue, black, white, gold, beige, or multicolor depending on what appears in the image.
Materials
Section titled “Materials”When the AI can identify a material from visual texture, it adds tags like leather, cotton, denim, metal, wood, silk, or plastic.
Scene context
Section titled “Scene context”The AI recognizes the type of setting the product was photographed in. Common scene tags include outdoor, studio, lifestyle, flat lay, on model, and close-up.
Background type
Section titled “Background type”Background tags help you filter images by how they were shot. Expect tags such as white background, transparent background, natural background, gradient, or textured background.
How to run auto-tagging
Section titled “How to run auto-tagging”Tag a single image
Section titled “Tag a single image”- Open your media library and click on the image you want to tag
- The preview panel opens on the right side of the screen
- Click the Generate Tags button below the image details
- Wait a few seconds while the AI analyzes the photo
- The generated tags appear in the Tags section of the preview panel
Tag multiple images
Section titled “Tag multiple images”There is no dedicated batch TAG button in the toolbar. To tag multiple images, open each file individually and use the Generate Tags button in the preview panel. This gives you the opportunity to review tags as they are generated for each image.
Reviewing and curating tags
Section titled “Reviewing and curating tags”Auto-tagging is a starting point, not the final word. The AI does the heavy lifting, but you stay in control of what tags each image carries.
Remove irrelevant tags
Section titled “Remove irrelevant tags”After the AI generates tags, review them in the preview panel. If a tag does not fit — for example, the AI tagged a navy item as black — click the X next to that tag to remove it.
Add custom tags the AI missed
Section titled “Add custom tags the AI missed”You can type any custom tag into the tag input field and press Enter to add it. This is useful for brand-specific terms, internal categories, or product attributes the AI cannot see, such as summer-2026 or best-seller.
Accuracy expectations
Section titled “Accuracy expectations”The AI performs best under conditions typical of professional e-commerce photography. Here is what to expect.
Best results with:
- Clear, well-lit product photos with minimal clutter
- Standard e-commerce setups (white background, studio lighting)
- Common product categories (apparel, accessories, footwear, home goods)
- Images where the product fills most of the frame
May need manual adjustment for:
- Unusual or niche products the AI has not seen many examples of
- Dark or heavily stylized photography
- Images with multiple products that share equal prominence
- Very small products photographed from a distance
As a general guideline, expect the AI to correctly identify the primary product type and dominant colors in the vast majority of well-shot photos. Material and scene tags are accurate most of the time but benefit from a quick review.
Token cost
Section titled “Token cost”Each image you auto-tag uses a small number of AI tokens from your account balance. The cost per image depends on the resolution and complexity of the photo, but a typical product image costs roughly one token per tagging operation.
You can check your current token balance and usage history on your account settings page.
Finding tagged images with search
Section titled “Finding tagged images with search”Once your images are tagged, the real payoff begins. Head to the media library search bar and type any tag value — leather, outdoor, red, shoe — and every image carrying that tag appears instantly.
You can also combine multiple tags to narrow results. Searching for dress white background finds only dress photos shot on a white background, filtering out lifestyle shots or other product types.
Next steps
Section titled “Next steps”- Manage and organize tags — Learn the workflow for reviewing, editing, and standardizing tags across your catalog
- Search and filter your library — Use tags and other filters to find exactly the images you need
- Models and pricing — Understand AI token costs and available models